Systems and methods using weighted-ensemble supervised-learning for automatic detection of ophthalmic disease from images

Abstract

Disclosed herein are systems, methods, and devices for classifying ophthalmic images according to disease type, state, and stage. The disclosed invention details systems, methods, and devices to perform the aforementioned classification based on weighted-linkage of an ensemble of machine learning models. In some parts, each model is trained on a training data set and tested on a test dataset. In other parts, the models are ranked based on classification performance, and model weights are assigned based on model rank. To classify an ophthalmic image, that image is presented to each model of the ensemble for classification, yielding a probabilistic classification score—of each model. Using the model weights, a weighted-average of the individual model-generated probabilistic scores is computed and used for the classification.

Description

PRIORITY INFORMATION[0001]This patent was filed under 35 USC 111(a) on the same day as U.S. patent application titled “Systems and Methods Using Weighted-Ensemble Supervised-Learning for Automatic Detection of Retinal Disease from Tomograms”, which by virtue of reference is entirely incorporated herein.FIELD OF THE INVENTION[0002]The present invention relates to automated detection of ophthalmic diseases from images of the eye and its parts.BACKGROUND OF THE INVENTION[0003]The eye is the primary sensory organ involved in vision. There are a myriad of diseases which can affect the eye and result in visual deficit or blindness. Some of these diseases, such as diabetes, are systemic conditions which result in multiple organ dysfunction. Others of these diseases, such as age-related macular degeneration and primary open angle glaucoma, are primarily localized to the eyes. There is a significant and growing shortage of trained eye care providers competent to diagnose both primarily ophth...

AI Technical Summary

Benefits of technology

The patent is about a method for assigning weights to different models that are used to diagnose disease. The method aims to not give too much importance to models that perform poorly in the testing environment, and to not give too little importance to models that perform well. The method can be applied to various ophthalmic imaging modalities, such as BSCAN ultrasounds, fundus photographs, OCT imaging, and computed tomograms (CT scans). The disclosed invention can utilize existing and future imaging modalities, and can adapt itself to changes in technology.

Problems solved by technology

Nonetheless, there is significant overlap between the utilities of the various modalities.

Images